The Effect of Electronic Word of Mouth on Movie Sales: Focusing on Time, Channel, and Value

Abstract

With the rapid spread of social media, numerous things never imagined before are being realized. Many changes are evident in every corner of society because of the effects of social media. The current situation indicates a recent upsurge of interest in eWOM from various social media channels. This study overcomes the limitations of the existing research that could not present a unified interpretation of electronic word of mouth (eWOM). Thus, this study aims to analyze the impact of eWOM from three aspects of eWOM, namely, the time when eWOM is written, the social media channel that eWOM is written, and the value of eWOM. Our detailed research questions are as follows:  Does the difference between pre- and post-consumption eWOM (disconfirmation) have an impact on movie revenue?  How does the impact of eWOM on movie revenue change depending on the difference of social media channels on which eWOM is written?  Is the impact of a valuable eWOM stronger than other forms of eWOM? If so, what is a valuable eWOM? This study collects eWOM information on movies from February to August 2012 from Twitter, Yahoo! Movies (online review site), blog, and YouTube (free video sharing site) on a daily basis from two or three weeks before opening to closing. This study also collects the sales information on the movies on a weekly basis from BoxOfficeMojo.com. Using text mining analysis, this study classifies tweets into four, namely, intention tweet, positive tweet, neutral tweet, and negative tweet. Additionally, to examine the value of eWOM, 15,059 online reviews on various products in Amazon.com are randomly collected. Based on these data, three studies are conducted to answer the research questions. The first study on time of eWOM investigates the impact of pre- and post-consumption eWOM on sales based on expectation confirmation theory. Consequently, the disconfirmation in the number of tweets before and after opening has a negative impact on the film sales. This premise suggests that forming a higher expectation than the substantial value does not always bring about positive results, in the current situation that the eWOM on the substantial value of movies spreads rapidly. The second study is a comparative analysis of eWOM channels on how the impact of eWOM on movie sales differs on the social media channel where eWOM is written. Based on Rogerss innovation diffusion model, this study analyzes whether the impact of each social media on innovators and imitators is significantly different. The study also analyzes whether the impact in the initial and in the late stage of opening is significantly different. The result indicates that Twitter is relatively influential on the innovators in the initial stage of opening because of its mass media characteristic of spreading information rapidly in real time. On the other hand, Yahoo! Movies, one of online review sites, is found to be relatively influential on the imitators in the late stage of opening because of its interpersonal communication characteristic, showing a strong persuasion by collective intelligence. This study is academically significant because this is the first work that attempts to classify social media channels from the perspective of the impact of social media using actual eWOM data. Moreover, this study is expected to be utilized as basic data to present strategic directions from the corporate perspective in which social media is utilized by period or by purpose. This study also suggests the future direction of social media evolution. The third study on the value of eWOM reveals that all forms of eWOM do not have the same influence, and determines which eWOM is valuable. The existing research could not draw consistent conclusions on the relationship between rating extremity and review helpfulness. The current study explores these reasons from the perspective of collective intelligence. Online review information from Amazon.com is utilized in this study, and the response surface methodology is used to demonstrate that when the individual review ratings are more consistent with the product average ratings, the helpfulness of reviews increases. In other words, most consumers tend to believe the reviews that are close to product average rating which is a kind of collective intelligence is more helpful. This observation can be interpreted to show that collective intelligence, which plays a major role in forming consumer attitudes toward product quality, is an important factor to determine review helpfulness. The findings of this study are expected to help draw a strategy that is available to reduce the information overload on consumer purchase decision and make online marketers understand the characteristics of helpful reviews. This study has three contributions. First, the study contributes from the research methodology perspective. For this analysis, eWOM is collected as panel data from Twitter, Yahoo!Movies, YouTube, and blog. Text mining is utilized to collect the information on the contents of eWOM as quantified variables. eWOM information from various social media channels is utilized, and a more extensive analysis is applied compared to the existing research. Second, the study contributes from the theoretical perspective. This study is significant in that the disconfirmation between pre- and post-consumption eWOM on the movie has a negative impact on the movie sales based on expectation confirmation theory. This study also determines that social media is divided into the media that has the characteristic of mass media or interpersonal communication based on the innovation diffusion model. Third, this study contributes from the practical perspective. The study is expected to present strategic directions for corporate eWOM management. Based on the findings from this study, a company is expected to identify eWOM characteristics by period and by media, and consider the differentiation of valuable eWOM that can be utilized to reveal more efficient eWOM management.